How Omnyx Works

Your AI executive assistant that plugs into your stack and runs your communication loop

Overview

Omnyx is your AI executive assistant that builds a long-term memory of your work on top of your communication stack. It connects to your email, calendar, phone, documents, and (optionally) social channels, and handles as much of your communication loop as you allow.

Think of it as your second brain: it learns how you think, write, and decide, then drafts replies, books meetings, answers calls, and follows up—all in your voice and tone. You stay in control through our Approvals system, or you can let it run on Autopilot for the routine stuff.

In practice, most founders start by letting Omnyx run their email and calendar, then turn on calls and social channels once they trust the clone and have seen it behave correctly with their executive workflows.

Everything lives inside a single web dashboard: a unified inbox showing all threads (email, call summaries, tasks), your Memory Engine vault, Approvals queue, Autopilot settings, and team/integration controls. No juggling tabs. No context-switching.

Onboarding Flow

Getting started with Omnyx takes about 5 minutes. Here's what the flow looks like:

1

Create your account

Choose your role (founder, CEO, executive, agency owner), region, and time zone. This helps Omnyx understand your context from day one.

2

Connect Gmail + Calendar

One-click OAuth flow from Settings. Omnyx only takes actions you explicitly allow—you decide what it can read, draft, and send.

3

Define your persona & tone

Quick questionnaire: How do you like to sound in emails and calls? What do you handle personally vs. delegate? This trains your digital clone.

4

Seed your Memory Engine

Upload a few key documents (playbooks, FAQs, sales decks). Connect folders like "Clients" or "Investor Updates". Omnyx uses these to ground every reply.

5

Choose Autopilot defaults

Decide where Omnyx should only draft (Approvals mode) vs. where it can send/book without asking (Autopilot mode). You can change this anytime.

Dashboard & Inbox

Your Omnyx dashboard is command central. The main view shows a unified inbox of all threads—emails, call summaries, and tasks—all in one feed.

Each thread clearly shows:

  • The original message or call transcript
  • Omnyx's proposed reply or action
  • The context it used (docs, prior emails, your calendar)
  • One-click Approve/Send or Edit/Override buttons

Autopilot items (already handled by Omnyx) are marked with a green "Handled" badge. Approvals (waiting on you) are marked "Pending Review". Everything is searchable and filterable by channel, date, or contact.

Unified Inbox
Handled
Pending
From: client@example.com
Re: Project timeline update
Omnyx drafted: "Thanks for the update..."
Approve
Edit

Email & Calendar

Omnyx integrates with Gmail and Google Calendar to handle your email and scheduling. Once connected, it can:

  • Draft replies to incoming emails in your voice and tone
  • Propose meeting times based on your availability
  • Create, update, and cancel events on your calendar
  • Follow up with people who didn't respond to invites

In Approvals mode (the default), Omnyx will never send an email or book a meeting without your explicit approval. Every draft appears in your inbox with a one-click "Approve & Send" button.

In Autopilot mode, you can let Omnyx send and schedule directly, within limits you define (e.g., "Book follow-ups under $5k automatically, but not investor calls"). You stay in control.

Sample Flow:

1
Inbound email from prospect
2
Omnyx drafts reply using your playbook
3
You approve in dashboard
Omnyx sends & marks thread as handled

Zoom Meetings

Connect your Zoom account and Omnyx becomes your meeting memory. Every call is automatically captured and processed:

  • Automatic transcription of all your Zoom calls
  • AI-generated summaries with key decisions and action items
  • Contact enrichment: Omnyx learns about the people you meet
  • Follow-up drafts: Auto-generated email drafts based on what was discussed

After each meeting, Omnyx extracts commitments, decisions, and next steps—then links them to the relevant contacts and accounts in your Memory Engine.

🎁 Founding Customer: Unlimited Zoom transcription minutes included at $19.99/mo. Lock your rate before public launch at $500/mo.

Social Channels

LinkedIn is live and production-ready. Connect from Settings → Integrations to let Omnyx:

  • Draft posts in your voice and style
  • Handle comments and engagement
  • Manage DMs with the same Approvals or Autopilot logic

All channels share the same Memory Engine, so Omnyx knows your brand voice, key talking points, and what you've said before.

LinkedIn
Available
X (Twitter)
Coming Soon
Instagram
Coming Soon
Reddit
Coming Soon

Memory Engine

The Omnyx Memory Engine is what makes your assistant actually remember. Unlike chatbots that start from zero every session, Omnyx builds a long-term memory from your work—so answers get sharper over time instead of staying generic.

What Gets Remembered

  • Documents: Every PDF, deck, playbook, and file you upload
  • Emails & threads: Context from important conversations, not just single messages
  • Meetings: Zoom calls, transcripts, decisions, and action items
  • Promises & Commitments: What you told investors, customers, and team
  • People & Accounts: Profiles of who you work with and their preferences

How It Helps You

  • "What did I promise Acme last quarter?" — Omnyx retrieves the exact commitment with source citation
  • "Draft a follow-up for Sarah" — Uses context from past emails, meetings, and docs
  • "Summarize all customer feedback on pricing" — Aggregates insights across channels
  • "Who at Sequoia cares about enterprise?" — Pulls from meeting notes and email history

Gets Smarter With Use

The more you use Omnyx, the more context it has:

  • Week 1: Basic document retrieval and email drafting
  • Month 1: Knows your key accounts, main contacts, and recurring topics
  • Month 3+: Connects dots across channels—linking that Slack message to that meeting to that proposal

Privacy & Security: Your memory is encrypted and completely isolated—only you can access it. We never use your data to train AI models. You can delete any document or memory at any time, and all derived data is purged with it.

🎁 Founding Customer: Unlimited memory vault + full API access included at $19.99/mo (locked for 9 months). Public launch price = $500/mo.

Approvals vs Autopilot

Omnyx gives you two modes for every channel and use case:

Approvals (Default)

Omnyx prepares drafts—emails, calendar invites, social posts—and waits for your approval. Nothing goes out without your explicit click. This is the safest mode and the default for sensitive channels.

Autopilot (Opt-in)

You define policies where Omnyx can send or book without asking. Examples:

  • "Follow up with leads under $5k on Autopilot"
  • "Book client calls automatically, but not investor calls"
  • "Reply to support emails tagged 'FAQ' on Autopilot"

Configure this from Settings → Automation. Your inbox clearly labels which actions were Autopilot (green "Handled" badge) vs. which are waiting for you (yellow "Pending Review").

Settings → Automation

Email - Approvals mode
Calendar - Autopilot for routine bookings
Calls - Autopilot for screening

Security & Permissions

Once you connect your accounts, here's what Omnyx can see and do:

Email (Gmail)

  • Read messages in your connected account
  • Draft and send replies (only when you've allowed it via Approvals or Autopilot)
  • Never access emails you explicitly exclude or mark as private

Calendar (Google Calendar)

  • View your calendar to check availability
  • Create, move, and cancel events (only what you approve or allow via Autopilot)
  • Respect your working hours and time zone settings

Documents & Memory

  • Read only what you upload or connect (Google Drive folders, etc.)
  • Store encrypted chunks in your workspace's secure memory vault
  • Never use your data to train public models

What Omnyx does NOT do:

  • No cross-tenant sharing (your data stays in your workspace)
  • No training public models with your content
  • No selling or sharing your data with third parties

All data is encrypted at rest and in transit using industry-standard encryption. You can revoke access anytime from Settings → Integrations.

Developer API

For developers and power users, Omnyx exposes a RESTful API so you can integrate the Memory Engine into your own applications, workflows, and automations.

API Keys

Generate API keys from your Dashboard → Developer Console:

  • Scoped permissions: Control access per key (documents, memories, search, router, connectors, admin)
  • Usage tracking: See when keys were last used and revoke anytime
  • Multiple keys: Create separate keys for different integrations or environments
  • Secure by default: Keys are hashed and never stored in plaintext

Available Endpoints

POST /api/memory/documents
Upload documents (PDF, DOCX, TXT) to your Memory Engine. Content is chunked, embedded, and indexed automatically.
POST /api/memory/memories
Create entity memories (facts about contacts, accounts, or custom entities). Supports static facts and learned behaviors.
POST /api/memory/search
Semantic search across your documents and memories. Returns ranked results with similarity scores and source citations.
POST /api/router/chat
Memory-augmented chat completions. OpenAI-compatible format with automatic RAG context injection.
GET /api/memory/documents
List all documents in your memory vault with pagination, filtering, and metadata.

Example: Memory-Augmented Chat

# Ask a question with automatic memory context
curl -X POST https://api.omnyxai.com/api/router/chat \
-H "Authorization: Bearer $OMNYX_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-4o-mini",
"messages": [{"role": "user", "content": "What did we promise Acme Corp?"}]
}'

Use Cases

  • Custom integrations: Connect Omnyx to Slack, Discord, or your own apps
  • Workflow automation: Trigger memory searches from Zapier, Make, or n8n
  • Internal tools: Build custom dashboards powered by your memory vault
  • AI agents: Give your agents access to long-term memory and context

🎁 Founding Customer: Full API access included at $19.99/mo. Interactive documentation and code samples available in the Developer Console.

14-day free trial • No credit card required